poly(vinyl chloride) and wood flour press mould composites: new bonding strategies

Poly(vinyl Chloride) and Wood Flour Press MouldComposites: New Bonding Strategies

Nuno Rocha,1,2 Jorge F.J. Coelho,1 Ana C. Fonseca,1 Algy Kazlauciunas,2 Maria H. Gil,1

Pedro M. Goncalves,3 James T. Guthrie2

1CIEPQPF, Chemical Engineering Department, University of Coimbra, Coimbra 3030-790, Portugal2Department of Colour Science, University of Leeds, Leeds LS2 9JT, United Kingdom3Cires S.A. – Companhia Industrial de Resinas Sinteticas, Apartado 20, Estarreja 3864-752, Portugal

Received 19 September 2008; accepted 15 March 2009DOI 10.1002/app.30419Published online 29 April 2009 in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: In this work, different strategies forimproving the association between hydrophilic wood floursurfaces and poly(vinyl chloride) (PVC) hydrophobicsurfaces were tested. Three new coupling agents, based onliving radical polymerisation (LRP), involving PVC weresynthesised and tested in formulations with PVC andwood flour. The melt mixing behaviour was analysed interms of the torque exerted by the mixing blades andrelated to the structural properties of the mixture. Theseproducts were ground and sheets were produced by pressmoulding. The composites were characterised by dynamic

mechanical analysis. It was found that the use of a newblock copolymer poly(vinyl chloride)-b-poly(hydroxy-propyl acrylate)-b-poly(vinyl chloride), prepared by LRP,increases the elastic modulus of the composite, under con-trolled conditions, involving the use of specific amounts ofthe copolymer. VVC 2009 Wiley Periodicals, Inc. J Appl Polym Sci113: 2727–2738, 2009

Key words: poly(vinyl chloride) (PVC); composites; livingpolymerisation; rheology; thermal properties

INTRODUCTION

Poly(vinyl chloride) (PVC)/wood flour compositesare the subject of considerable research. These mate-rials are used in various applications such as win-dow/door profiles, decking, railings, and sidings.Wood flour polymer composites can have a longerlife than native wood, as they are less susceptible tomoisture-induced swelling, shrinking, and biodeter-ioration. Wood plastic composites (WPC) takeadvantage of the beneficial properties of both thewood and the plastic components.1–6

The main disadvantage with most composites ofthis type is the weak interfacial adhesion that arisesbetween the two components. A possible solution isto use compatibilisers, such as block copolymers andcoupling agents. The inclusion of the most commonmaterials provides increasing hydrophobicity in thewood flour phase, due to chemical reaction of eachspecific group with the hydroxyl groups of thewood. The surface tension of wood flour is reducedand approaches that of the molten thermoplastic.The most common coupling agents are organicacids, anhydrides, isocyanates, and silanes.7–12

The use of silane-based coupling agents offers oneof the more successful routes to improving the asso-ciation between fibres and polymers. These can actas a bridge between the hydrophilic regions and thehydrophobic regions of the materials. Acids andbases catalyse the hydrolysis of the alkoxy silane tothe corresponding silanol, which can later undergocondensation reactions. However, the formation ofpolysiloxanes, inhibits the adsorption of the silanebeyond the fibre surface, so that in highly concen-trated silane solutions, there is no additional silaneadsorption. After hydrolysis, the silanols slowly con-dense to form oligomeric species.13–19

The conventional free-radical polymerisation ofvinyl chloride is the only industrial process that isused for the synthesis of PVC. Living radical poly-merisation (LRP) provides an efficient tool for poly-mer synthesis, yielding macromolecules withcontrolled molecular weights and narrow polydis-persities. The macromolecules are also free of struc-tural defects and contain long-lived chains that canbe further activated. The LRP of VC, initiated withCHI3 and catalysed by Na2S2O4, is mediated mainlyby a combination of competitive single electrontransfer (SET) and degenerative chain transfer(DT) and provides a novel pathway to living radicalpolymerisation. This reaction can take place in anaqueous medium at room temperature. This poly-merisation process is technologically important

Journal ofAppliedPolymerScience,Vol. 113, 2727–2738 (2009)VVC 2009 Wiley Periodicals, Inc.

Correspondence to: N. Rocha ([email protected]).

because it requires only common industrial facilitiesto produce PVC materials that possess the propertiesdescribed above.20–24

The chloroiodomethyl chain ends, arising from theLRP-PVC, can be replaced with other reactivegroups, acting as a macroinitiator. The functionalisa-tion of the chain ends is accessible by SET-mediatedorganic reactions. This method provides the firstexample of telechelic a-b-di(hydroxy)PVC.23

Copolymers have been added to improve the me-chanical properties of WPC. This improvement isdue to a better dispersion of fibres in the matrix, amore effective wetting of the fibres in the polymericmatrix and better association between the twophases. The grafting of the wood also presents apromising approach to the modification of thesematerials, by the inclusion of hydrophobic groupsinto the wood flour surface.25–27

The ability to synthesise block copolymers of PVCwith acrylates, using the single-electron-transfer/degenerative-chain-transfer LRP (SET-DTLRP) tech-nique is extremely important because, by thismethod, it is possible to control the influence of thepolymers with their different nature on the PVCproperties.28–31

Hydroxyl groups can be incorporated in a con-trolled way into the PVC, by creating a block copoly-mer of PVC with poly(hydroxypropyl acrylate)(PHPA).32 A new pathway in WPC is thus createdsince the interaction is increased by modification ofthe plastic phase, increasing its hydrophilicity. Thisopens up the possibility for the PVC to bond with thewood flour by hydrogen bonding, or crosslinking,thereby decreasing the degree of wood flour self-entanglement. In the case of PVC-b-PHPA-b-PVC, thecopolymer would be expected to interact with thePVC, via the PVC segments, and with the wood flour,via the hydroxyl groups.

This study was aimed at the synthesis of new cou-pling agents for WPC, based on LRPPVC, to enablecoupling between the PVC and the wood flour thatdisplayed strong intermolecular forces via hydrogenand covalent bonding. The materials obtained werecharacterised by melt mixing torque rheometry andby dynamical mechanical thermal analysis.

EXPERIMENTAL

Materials

Wood flour (WF) from Bubinga (Switenia macro-phylla) was donated by DPM - Distribuicao e Produ-cao de Moveis (Portugal). These fibres were groundin a cutting mill, Retsch GmbH SM1, with a 500 lmstainless steel trapezium shaped sieve.

PVC (a suspension grade, with a molecular weightof about 63,000 g mol�1, Kvalue¼ 65.88, and d50 ¼

168.2 lm) was supplied by Cires, SA (Estarreja,Portugal).Stabilox CZ 2973 GN (calcium and zinc based

soap stabilizer and lubricant), sodium stearate (lubri-cant), epoxydised soya bean oil (co-stabiliser and in-ternal lubricant in PVC formulation), Kane AcePA210 (acrylic copolymer (MD-P210-C210) impactmodifier and lubricant), and wax PE520 (KWPEI 15)(lubricant for plastics) were provided by Cires.Hydroxypropyl acrylate (HPA) (95%), dimethyl

sulfoxide (99þ%), trimethoxyvinylsilane (97%), so-dium dithionite (85%) (Na2S2O4), sodium bicarbon-ate (99þ%) (NaHCO3), 2-allyloxyethanol (98%), andiodoform (99%) (CHI3) were purchased from Sigma-Aldrich (Portugal).Isopropanol, hydrochloric acid (37% w/w), etha-

nol and methanol were purchased from Jose M. VazPereira, S.A. (Portugal). Hydroxypropyl methylcellulose (Methocel F50) and partially hydrolysedpoly(vinyl alcohol) (PVA) were provided by Cires,S.A (Portugal). Vinyl chloride was purchased fromShinEtsu (Japan).

Coupling agents

LRP-PVC preparation

The LRP-PVC synthesis was carried out in a 5L pilotreactor charged with 3.00 L of water, 128.84 g of a3% aqueous solution of PVA, and 89.06 g of a 1.86%aqueous solution of methocel MF50. The other com-pounds were then weighed and charged into the re-actor: 9.24 g of CHI3, 32.70 g of catalyst Na2S2O4,7.89 g of buffer NaHCO3, and 1.00 L of vinyl chlo-ride monomer (VCM). The reactor was closed andsealed. The oxygen was removed from the reactorby applying four cycles of vaccum and nitrogen. TheVCM (1.00 L) was charged and the reactor washeated up to the polymerisation temperature (42�C).The reaction time was 24 h. The product was rinsedthree times with water and then filtered. The LRP-PVC was dried in an oven at 70�C.

Wood flour-PVC covalent bonding (PVC-WF)

A solution containing 225 mL of trimethoxyvinylsi-lane, 50 mL of distilled water, 240 mL of isopropa-nol, and 3 mL of chloridric acid was prepared. 8.5grams of Bubinga wood flour were poured into theprepared solution. The contents were left in a 1-LParr Instrument Co. 4523 Pressure Reactor for 5 h at30�C, with an agitation of 200 rpm and with anabsolute pressure of 3 bar. The fibres were washed 5times with distilled water using a vacuum filtrationunit. The modified fibres (WFmod) were dried atroom under vacuum for 4 days at 70�C.

2728 ROCHA ET AL.

Journal of Applied Polymer Science DOI 10.1002/app

A 1-L Parr Instrument Co. 4523 Pressure Reactorwas charged with 400 mL of dimethyl sulfoxide,4.400 g of LRP-PVC [Mn (TriSEC) ¼ 41,196 g mol�1],41.7 mg of sodium dithionite, 14.8 mg of sodium bi-carbonate (99þ%) and 1.500 g of WFmod. The reac-tor was conditioned with nitrogen for 10 min. Thenthe inlet and outlet nitrogen valves were closed untilthe system reached a constant pressure of 2.5 bar.The reaction mixture was stirred at 200 rpm at 70�C.After 4 h, the reactor was emptied and the mixturewas poured into 3 L of ethanol. The product was fil-tered and rinsed three times with methanol anddried in a vacuum oven at 70�C.

LRP-PVC functionalisation (PVC-OH)

A 1-L Parr Instrument Co. 4523 Pressure Reactorwas charged with 600 mL of dimethyl sulfoxide,6.600 g of LRP-PVC [Mn (TriSEC) ¼ 41,196 g mol�1],1.400 g of Na2S2O4, 1.400 g of NaHCO3 and 24 mLof 2-allyloxyethanol. The reactor operation and prod-uct purification were undertaken in the same way aspreviously outlined in Section 2.2.1.

PVC-b-PHPA-b-PVC synthesis (PVC-PHPA)

The a,x-diiodopoly(HPA) synthesis was carried out ina 5 L pilot reactor by the SET-DTLRP route, initiatedwith 5.32 g of CHI3 in 1.2L deionised water at 25�C.The reactor was also charged with 0.40 L of HPA,9.42 g of catalyst Na2S2O4, 4.54 g of buffer NaHCO3,45.99 g of a 3% aqueous solution of PVA, and 31.79 gof a 1.86% aqueous solution of methocel F50. The reac-tion was undertaken during 6 h at a polymerisationtemperature of 42�C. A sample of the macroinitiatorPHPA was collected at the end of this first stage.

Afterwards, the reactor was charged with 1.5 L ofwater, 96.59 g of a 3% aqueous solution of PVA, 66.67g of a 1.86% aqueous solution of MF50, 18.84 g of a cat-alyst, Na2S2O4, 1.65 g of the buffer NaHCO3, and 0.75L of vinyl chloride monomer (VCM). The reactor washeated up to the polymerisation temperature (42�C)and the reaction was undertaken during 24 h. This laststage makes possible the completion of the synthesisof the copolymer PVC-b-PHPA-b-PVC (PVC-PHPA).

Before each of the two reactions, the reactor wasclosed and flushed 5 times with nitrogen for 5 min.A vacuum was applied until 0.10 bar for 5 min. The

product was rinsed 3 times filtered with water anddialysed in water. Finally, the block copolymer wasdried in a vacuum oven at 70�C.

Processing

Preparation of PVC – rigid formulation

PVC and the additives needed to create a rigid stand-ard formulation were mixed in a high-speed mixer.The solid components were poured into the mixerand mixed for 1 min at 1000 rpm. Then the speedwas increased to 2500 rpm. This led to a temperatureincrease and when it reached 70�C, the speed wasdecreased to 1000 rpm and the liquid componentswere added. Then the speed was increased again to2000 rpm. Finally, when the temperature reached110�C, the cooling water was turned on and thespeed reduced to 1200 rpm. The mixture was col-lected when the stock temperature was about 50�C.The PVC rigid formulation was prepared with 100

phr (per hundred of PVC) of PVC, 4 phr of StabiloxCZ 2973 GN, 3 phr of epoxydised soya bean oil, 0.5phr of sodium stearate, 1 phr of PA210, and 0.5 phrof PE520.

PVC, Bubinga wood flour and coupling agentsblending

The composites were compounded using a Plasto-graph Brabender mixer W50EHT, with two rotors. Themixing temperature was set to 200�C. The mixture wascollected after it reached the melting point, to avoiddegradation. The fusion chamber was loaded with40 grams of raw-materials and the mixing speed set to40 rpm. Table I gives the composition of each sample.

Press mould composites

The fused and mixed materials, obtained in a solidform were then ground in a Retsch GmbH SM1 cut-ting mill, fitted with a 500 lm stainless steel trape-zium shaped sieve.The obtained ground materials were then used to

manufacture the desired boards by press mouldingin a laboratory hydraulic hot press (412BCE, Carver).The materials were pressed for 2 min at 4 metrictons, to reach the plates’ temperature (190�C). Theywere then pressed for 3 min at 11 metric tons. After-wards, cooling was applied with pressurised airuntil a temperature below 100�C was reached.

Characterisation techniques

Dynamic mechanical thermal analysis

Dynamical Mechanical Thermal Analysis (DMA) of2 mm thick specimens, of a size 6.5 mm � 6.0 mm,

TABLE IComposition of Each Mixture (% w/w)

Component M00 M01 M02 M03 M04 M05

PVC 100.0 80.0 70.0 70.0 76.5 76.5WF 20.0 20.0 20.0 20.0 20.0PVC-OH 10PVC-PHPA 10 3.5PVC-WF 3.5

POLY(VINYL CHLORIDE) AND WOOD FLOUR PRESS MOULD 2729


was performed using a Triton Tritec 2000 unit in sin-gle cantilever bending, in the multifrequency mode(1 Hz, 3 Hz, 5 Hz ,and 10 Hz), with a standard heat-ing rate of 2�C min�1. The composites were previ-ously moulded with the required dimensions,following the procedure given in Section 2.3.3. Thepowders were characterised in the Constrain LayerDamping mode, with a standard heating rate of 2�Cmin�1, at the same frequencies as those indicatedabove. The glass transition temperature (Tg) wasdetermined as the peak in tan d (E00/E0). Here, E00

and E0 are the loss modulus and storage modulusvalues, respectively. Each run was repeated twice toestablish the reproducibility of the results. Theresults presented correspond to the average value.

Thermogravimetric analysis

Thermogravimetric Analysis (TGA) was used todetermine the thermal behaviour of the raw-materi-als and of the composites under non-isothermal con-ditions. The measurements were carried out using aTA Instruments SDT Q600 unit (thermobalance sen-sitivity: 0.1 lg). Alumina crucibles were employed.The calibration was performed in the range from25�C to 1000�C with tin and lead as the standards.Samples were loaded into open alumina crucibles. Adry nitrogen purge flow of 100 mL min�1 was usedin all measurements. Sample weights ranging from 5to 10 mg were used.

Nuclear magnetic resonance (NMR)

The 1H-NMR spectra (500 MHz) were recorded on aBruker DRX 500 spectrophotometer at 32�C in THF-d8.

13C NMR spectra of the solid samples wereobtained in a 9.4 T Bruker Avance 400 WB spectro-photometer (DSX model) on a 4 mm double-bearingprobe, at 100.6 MHz. The solid state NMR experi-ments were conducted at room temperature under aCPMAS frequency of 9000 Hz.

Fourier transform infrared spectroscopy

Fourier Transform Infrared Spectra (FTIR) wereobtained using a Nicolet 750 unit, with a GoldenGate ATR accessory (Specac). The products wereanalysed as prepared. The resolution was 4 and thenumber of scans was 64.

Size exclusion chromatography

The chromatography parameters of the sampleswere determined using a HPSEC unit (High Per-formace Size Exclusion Chromatography); Viscotek(Dual detector 270, Viscotek, Houston) with differen-tial viscometry (DV); right-angle laser light scatter-

ing (RALLS, Viscotek) and RI (Knauer K-2301)detection. The column set consisted of a PL 10-lmguard column (50 � 7.5 mm2) followed by twoMIXED-B PL columns (300 � 7.5 mm2, 10 lm). TheHPLC pump (Knauer K-1001) was set with a flowrate of 1 mL/min. The eluent (THF) was previouslyfiltered through a 0.2 lm filter. The system was alsoequipped with a Knauer online degasser. The testswere done at 30�C using an Elder CH-150 heater.Before injection (100 lL), the samples were filteredthrough a PTFE membrane with 0.2 lm pores. Thesystem was calibrated with narrow molar masspoly(styrene) standards.

Chlorine microanalysis

The determination of the chlorine content was per-formed by Schoniger flask combustion and titrationwith mercury nitrate. The sample (placed in plati-num basket) was combusted in a hermetically sealedSchoniger flask, previously filled with oxygen andallowed to stand for 20 to 30 min. Following thistime period, the basket was placed in the flask andwashed with the minimum amount of water. Brom-phenol blue and nitric acid solution were addeddropwise until yellow colour was achieved. Then 0.5mL more of nitric acid, 50 mL of ethanol, and di-phenyl carbazone indicator were added. Titration toa purple end point was undertaken with the 0.05 Mmercuric nitrate solution. A standard determinationwas carried out on a standard of known percentageof chlorine (p-chlorobenzoic acid).

Flexural experiments

The flexural tests were performed using a DARTECUniversal Testing unit with a servo-hydraulic func-tion of 200 kN capacity, fitted with a 5 kN loadcellcalibration, including increments of 0.001%. Thesetests were performed according to the ASTM D790-03 (procedure A), with five specimens for each sam-ple being tested (2.0 mm in thickness, 55 mm inlength, and 12.7 mm in width), a crosshead motionrate of 1.2 mm/min, until a midspan deflection of 6mm and a support span of 38 mm were achieved.

Scanning electron microscopy

The morphology of the materials was investigatedusing a JEOL JSM-820 scanning electron microscope(SEM). All of the specimens were coated with goldusing a Bio-Rad SC500 diode sputter coating unit.The samples were examined over the magnificationrange from �20 to �1000 using an accelerating volt-age of 5 kV.

2730 ROCHA ET AL.


RESULTS AND DISCUSSION

Coupling agents synthesis

The synthesis of PVC-WF and PVC-OH dependedon the synthesis of LRP-PVC. This polymer was syn-thesised through a mechanism that was developedin previous work.21 The proposed chemical structureis shown in Figure 1.

The LRP-PVC was analysed by 1H-NMR, revealingthe presence of terminal active chains ends (-CHClI)between 6.1 and 6.3 ppm and the absence of anystructural defects (Fig. 2) as described previously.32

SEC analysis revealed a numerical molecular weightof 42,000 g/mol with a polydispersivity of 2.831.

PVC-WF

During the first stage, the presence of water, chlori-dric acid, and isopropanol allows the hydroxylgroups (present in the cellulose of the wood) to reactwith the trimethoxyvinylsilane through partial hy-drolysis and co-condensation,17 making it possible tosynthesise WFmod. A possible chemical structurefor WFmod is shown in Figure 3.

The presence of the vinyl groups in WFmodallows functionalisation of LRP-PVC, by a mecha-

nism described in previous work.23 The doublebonds of WFmod react at the edges of the LRP-PVCthrough the iodine groups, resulting in a PVC-WF,with the chemical structure shown in Figure 4.Figure 5 represents the FTIR analysis of Bubinga

wood flour (WF), the Bubinga wood flour modifiedwith trimethoxyvinylsilane (WFmod), and the prod-uct of reaction between WFmod and LRP-PVC(PVC-WF). The WFmod trace shows that the vinylgroups were present in the sample. The PVC-WFspectra shows that significant proportion of the vinylgroups reacted with the LRP-PVC, as observed fromthe decrease in the appropriate FTIR band. Thebroad band between 1100 and 1000 cm�1 and thebands between 1000 and 950 cm�1 are due to thepresence of the -Si-O-Si- and -Si-O-CH2-bonds

32

respectively. Thus, Figure 5 shows the presence ofthese groups in WFmod and PVC-WF. This band isless pronounced in the spectrum of PVC-WF than inWFmod, indicating that the reaction with LRP-PVChas broken much of the associations between the sil-ane groups. The peak at 600 cm�1 indicates the pres-ence of C-Cl bonds which supports the presence ofPVC in the final structure.

PVC-OH

This telechelic material was functionalised followingthe procedure described in Section 2.2.3 for whichthe mechanism is described in previous work.23 Fig-ure 6, shows the chemical structure of PVC-OH.Figure 7 shows the FTIR spectra for the functional-

ised PVC and the comparison spectra of the raw-

Figure 1 Chemical structure of LRP-PVC21.

Figure 2 500 MHz 1H-NMR of the a,x-diiodo(PVC) pre-pared by SET-DTLRP.

Figure 3 Chemical structure of WFmod.

Figure 4 Chemical structure of PVC-WF.



materials. Hydroxyl groups were incorporated intothe PVC, as suggested by the presence of the broadband for hydroxyl group as a small between 3600and 3200 cm�1. Since the proportion of 2-allyloxye-thanol is very small compared with that of the LRP-PVC, it is not surprising to see that there is only asmall peak for the hydroxyl band in the FTIR spec-trum of PVC-OH. Note also, that the vinyl groupband at 1600 cm�1, of 2-allyloxyethanol, is not pres-ent for PVC-OH, suggesting that there is no residualmonomer present.

PVC-PHPA

As described in previous work,29 PVC-PHPA syn-thesis by the SET-DTLRP approach is achieved bysynthesising as the first stage the acrylate polymer,in this case poly(hydroxypropyl acrylate). In a sec-ond stage, polymerisation of VCM on the acrylatepolymer chain edges is brought about. The resultingstructure is represented in Figure 8.

Figure 9 shows the FTIR spectra for sample ofPVC-PHPA, its correspondent homopolymers andthe initiator iodoform. The copolymer demonstratesfeatures of both of the homopolymers, as well asthose of C-I bond active chain ends. The presence ofthe PHPA functional groups is shown by the broadband that corresponds to alcohol groups and estergroups. Carbon-chlorine bonds are not clearlyshown in this spectrum due to the lower level ofincorporation of PVC.

For a better evaluation of the inclusion of the PVCsegments (PHPA and PVC-PHPA being insoluble incommonly used solvents in NMR analysis), 13CNMR solid state analysis was undertaken on sam-ples PHPA and PVC-PHPA. The obtained spectraare presented in Figure 10. The presence of the PVCsegment of PVC-PHPA becomes clearer from thepresence of the peak at 58.7 ppm.34 The other PVCgroup that should be around 47.4 ppm34 is seem-ingly hidden by the PHPA methylene group peaks.The methyl group shows resonance signals between20 ppm, and the methyleneoxy groups, from the twoisomers of PHPA, show peaks between 63 and73 ppm.35 Carbonyl carbon atoms from PHPA canbe identified in the region between 170 and180 ppm.35,36 The active chain ends can be identifiedby the peak at around 17 ppm.37 The reductionof the peak from PHPA to PVC-PHPA can beexplained by the growth of the copolymer molecularweight through reaction of the former active chainends with VCM. Chlorine microanalysis indicates atotal PVC incorporation of 10.2 % (wt.) in the pro-duced copolymer.

Melt mixing behaviour

Figure 11 shows the torque that is exerted by theblades during the melt mixing for products on dif-ferent coupling agents and different compositions inthe polymeric phase. To avoid PVC degradation, the

Figure 5 FTIR spectra of WF, WFmod, PVC-WF, andLRP-PVC (vinyl, silane, and C-Cl regions).

Figure 6 Chemical structure of PVC-OH22.

Figure 7 FTIR spectra of PVC-OH, LRP-PVC and 2-ally-loxyethanol (hydroxyl and vinyl group regions).

2732 ROCHA ET AL.


mixing process was stopped at the first hint of mate-rial fusion.

The presence of wood flour strongly increases thetorque of the mixture. PVC powders and wood flourare very different in respect of the particle shapeand intermolecular forces. The increasing torque,when compared with that associated with the melt-ing of PVC powders, is attributed to such differen-ces. The inclusion of materials in the formulation,that can promote association between the PVC andthe wood flour phases, leads to a decrease in the tor-que. This effect is greater when PVC-PHPA is used.Even for lower levels than that which is used withthe other coupling agents, the effect of stabilising themixture (leading to a lower torque) is enhanced withthe increased PVC-PHPA content in the formulation.The presence of such a coupling agent allows for anincreased degree of entanglement between the PVCand the wood flour. This leads to a less effort beingrequired to promote their mixing.

Mechanical and thermal analysis

The mechanical and thermal characteristics wereevaluated by DMA and by TGA. DMA is probablythe most versatile thermal analysis method available.No other single test method providing more infor-mation from a single test.38 In DMA, the material issubjected to a forced mechanical vibration at a fixedfrequency, temperature, and elongation. A fractionof the energy is absorbed (E00, viscous component)and a fraction is returned elastically (E0, elastic com-ponent). The determination of the ratio of theseevents enables clear evaluation of the mechanicalperformance. Other parameters may also be moni-tored, such as the miscibility between different com-pounds in the composite and any thermal events.DMA, in multifrequency analysis, allows one to dis-tinguish between frequency dependent processesand those that are not. Molecular relaxations, such

Figure 8 Chemical structure of PVC-PHPA.

Figure 9 FTIR spectra of the PVC-PHPA, of the respec-tive homopolymers and of the initiator, iodoform.

Figure 10 13C Solid state NMR spectra of PHPA and ofPVC-PHPA.

Figure 11 Torque exerted by the mixing of the materialsin the fusion chamber, depending on the time and the cou-pling agents present in the formulation.



as a relaxations (or glass transitions), b relaxations,and the peak of tan d, are always frequency depend-ent.37 Examples of processes that are not frequencydependent are melting, thermal degradation, curingand crystallization. Figure 12 shows the DMA ther-mograms of the PVC that was used in this work.

Figure 12 reveals b and a transitions of the PVC at�49.5 �C and at 87.3�C, respectively. The tan d at 1Hz and the TGA traces of the major raw-materials,PVC and wood flour, are compared in Figure 13.

Figure 13 reveals two major thermal transitionsfor the wood flour that was used in this work. Thefirst transition is a broad peak in the DMA thermo-gram, between 10�C and 100�C, which can be relatedto the movement of chains that is associated to the

presence of moisture. The TGA thermogram of thesame sample shows that the first weight loss ends at100�C. The second peak, at around 240�C, is associ-ated with sample degradation and can be identifiedeither by TGA or by DMA. Regarding the PVC sam-ple, it is possible to observe frequency dependenttransitions (in Fig. 13 these are represented only thetan d at 1 Hz) that correspond to the b and a transi-tions of the PVC, respectively.32 According to thedegradation profile shown in Figure 13, the PVCand wood flour materials present higher degradationratios only after 240�C.Figure 14 illustrates the DMA and TGA traces for

M01 (Table I). Sample M01 is the reference material,used for thermal analysis since it contains only thePVC and the wood flour.The presented profile shows typical thermal

behaviour for a plastic material. E0 decreases nearlysteadily with the temperature until the glass transi-tion temperature (temperature at which the greatestdecrease in E0 occurs). The Tg, at 89�C, can be con-firmed by the sensitivity of tan d at this temperatureto the frequency of analysis. The tan d trace of M01is close to that of the PVC materials, in terms ofdamping peaks (for shape and temperature values,see also Fig. 12). This suggests that the wood flourplays the role of filler in the composite, the PVCbeing responsible for the matrix of the composite.After the Tg, the composite presents no plateau oflog E0 and loses very quickly its elastic capacity.Rubbery flow occurs just after the Tg. At around180�C, the composites become extremely soft andthe single cantilever experiment ends. At this tem-perature, the matrix made of PVC loses its mechani-cal resistance, reaching the softening temperature ofthe PVC. In terms of the degradation profile thecomposite shows a similar pattern to that observedwith PVC alone. It is interesting to note that theweight loss up to 100�C, for sample M01, is minimal

Figure 12 DMA of PVC (raw material).

Figure 13 DMA (at 1Hz) and TGA traces of PVC (rawmaterial) and wood flour.

Figure 14 DMA (at 1Hz) and TGA traces of M01 (Table I).

2734 ROCHA ET AL.


compared to the mass loss of the wood flour alone.The cause of this effect is the high processing tem-perature that eliminates the wood flour moisture.

The elastic modulus of the different composites ispresented in Figure 15. The results suggest that thedifferent composites present the same mechanicalbehaviour within the range of the testing tempera-tures. The different samples show a consistentdecrease in the elastic modulus with temperatureincrease until around 50�C. After this, a sharpdecrease occurs. This indicates the beginning of themovement of the main chains.

A closer view of the traces over very low tempera-tures is presented in Figure 16, showing clearly therelative values of the E0 for the different compositesat these lower temperatures.

The results show that the amount of VC-PHPAthat is added can have dissimilar effects on the elas-tic modulus. The addition of 3.5% to the formulationcreates a higher elastic modulus in the composite,whereas the addition of 10% creates a lower elasticmodulus. Increasing the amount of PVC-PHPA canlead to such a degree of association between thePHPA molecules that the elastic properties of thematerial become poorer. In spite of the dissimilarbehaviour described above, it is interesting to noticethat, at �20 �C, the elastic modulus of the majorityof the samples is similar. The only samples that falloutside this trend are the samples to which the co-polymer of PVC-PHPA was added. The mechanicalproperties of the composites, with increasing tem-perature, are affected by the presence of the woodflour and the coupling agents, leading to slight dif-ferences in the values of E0 for each temperature.The influence of PVC-OH and PVC-WF, used ascoupling agents, is less pronounced compared withtheir influence on the composites produced using

PVC-PHPA. Functionalised PVC-OH and PVC-WF,when used as coupling agents in the compositePVC/wood flour, did not introduce major differen-ces in the composites. However, considering theresults obtained, the addition of the copolymer PVC-b-PHPA-b-PVC may be a matter of optimisation, interms of mixing and composition, to achieve thedesired performance properties.The E00 obtained for the composites is presented in

Figure 17. Here it is possible to verify that the com-posite with the lower viscosity is the reference, pre-pared with only PVC and wood flour, confirmingthe success in the enhancement of the associationbetween the PVC and wood flour, when the cou-pling agents were added. It is interesting to note

Figure 16 Elastic modulus of the different compositesbelow 0�C.

Figure 17 Loss modulus of the different composites at1 Hz.

Figure 15 Elastic modulus of the different composites.



that the composite M04 (Table I) is the sample thatpresents the higher viscosity and elastic modulus.This result suggests the presence of increasingentanglement between the PVC and the wood flour.

Figures 18 and 19 show SEM electron micrographsof a composite that was formulated without cou-pling agents (M01) and one formulated with 3.5% ofPVC-PHPA (M04). For M01, it is possible to observeclearly the defined forms of the wood flour. How-ever, for sample M04, because of the coupling agent,and, in agreement with the DMA results, the com-patibility between the wood flour and the PVC ma-trix is increased. Thus, in Figure 19 it is no longerpossible to distinguish clearly the full wood flourparticles, due to a better entanglement with the PVCmatrix.

M02 (Table I) shows a profile that is slightly dif-ferent from what was expected, considering that theelastic modulus results did not show major differen-

ces from the results for M01. Two possible explana-tions can be given to this profile. One concerns theplasticising effect created by the inclusion ofhydroxyl groups at the LRP-PVC chain ends. Theseconds concerns of small amounts of the possibleremaining reaction solvent. However, it should benoted that the product was washed and dialysedseveral times before thorough drying to constantmass.Figure 20 presents the damping effect for the dif-

ferent composites that were prepared for this work.The results indicate that the Tg is only slightlyaffected over the different formulations studied. Theonly exception arises with sample M02 which has alower Tg than that of the other samples. The pres-ence of the small peak at �6.0�C should be noted.

Figure 20 Tan d of the different composites.

Figure 21 Tan d of the block copolymer PVC-b-PHPA-b-PVC.

Figure 19 SEM electron micrograph of sample M04 (Mag.�70).

Figure 18 SEM electron micrograph of sample M01 (Mag.�70).


2736 ROCHA ET AL.

Figure 21 represents the DMA plot of the block co-polymer PVC-b-PHPA-b-PVC, that was used to for-mulate the composites M03 and M04 (Table I). FromFigure 21 it is possible to identify the small peak at�6.0 �C as being the Tg for the block copolymer. Thepresence of this peak in the DMA trace of the com-posites M03 and M04 indicates some immiscibilityof PVC-PHPA. However, the immiscibility issuedoes not account for the increase of E0 for the com-posite M04, as seen in Figure 21. One possible rea-son for the presence of the immiscibility could bethe very high proportion of PHPA relative to thePVC in the copolymer (around 9 : 1 in weight). Thisleads to a copolymer that is rich in PHPA and witha greater tendency to stick together, causing phasesegregation during the processing.

The different composites present a small peak thatis sensitive to the applied frequency at temperaturesaround �50�C. This corresponds to the b transitionof PVC that did not disappear, regardless of the for-mulation used, confirming the high dependence onthe PVC matrix.

Table II summarises the major mechanical andthermal parameters for the different compositesstudied.

The composite M01 presents the lower values ofE0 compared with those for the other composites.The enhancement of the elastic modulus that wasobserved for the other composites of PVC/woodflour can be attributed to greater association betweenthe wood flour and the polymeric matrix, in thiscase the PVC. The presence of these coupling agentsmakes possible a real reinforcement of the PVC ma-trix with the wood flour. This increase in interfacialassociation, and in the degree of reinforcement,could also restrict the free motion of the PVC chainmatrix, leading to an increased viscosity and a corre-sponding increase in the loss of modulus,11 asobserved in the E00 values. The glass transition tem-peratures are slightly affected, but the presence ofthe coupling agents seems to give a plasticisingeffect to the product relative to the flexibility that isshown by the PVC/wood flour formulation withoutthe coupling agents. In terms of the energy involvedin the glass transition process, the values are in the

same range. However, the increasing complexity ofthe system seems to cause a moderate decrease inthese energies. Again, M02 presents values that arecompletely different.To validate the consistent results obtained from

the DMA analysis, composites containing 50% ofBubinga wood flour together with 0, 5, and 10% byweight of PVC-b-PHPA-b-PVC were submitted forflexural experiments. The strengths at yield pointwere 91.7 � 4.7, 103 � 1.0, and 77.5 � 5.8 MPa,respectively. These results are in agreement with thetrends shown in DMA results, indicating improvedmechanical properties for samples with smallamounts of PVC-b-PHPA-b-PVC and then a decreasein the strength of the composite as this amountincreases.

CONCLUSIONS

The results presented suggest the possibility forimproving the mechanical properties of PVC/woodflour composites by adding components that wereprepared using a living radical polymerisationmethod. These materials improve the melt mixingprocess, leading to a lower stress being applied dur-ing the mixing. The block copolymer PVC-b-PHPA-b-PVC is the coupling agent that gave the bestresults, leading to a mixing process for the PVC andthe wood flour that is similar to that of the PVCalone. This block copolymer, at certain loadings,enhanced the mechanical properties of the PVC/wood flour composites. However, for higher load-ings phase separation, due to increased associationbetween PHPA molecules, can occur leading topoorer mechanical properties. The results showedthe improved reinforcement of the PVC matrix bythe wood flour, when coupling agents, of the typedescribed, are used.

References

1. Jain, R. K.; Singh, Y.; Rai, M. Building Environ 1977, 12, 277.2. Stokke, D. D.; Gardner, D. J. J Vinyl Additive Technol 2003, 9,

96.3. Jiang, H.; Kamdem, D. P. J Vinyl Additive Technol 2004, 10, 59.

TABLE IIMechanical and Thermal Properties of the Composites

Sample

E0 (GPa) E0 (GPa)

Tg (�C) DH (kJ/mol)�50�C 0�C �50�C 0�C

PVC 2.10 1.80 82 46 87.3 242.5M01 2.86 2.38 82 46 89.0 210.6M02 2.76 2.36 71 53 81.1 11.5M03 2.60 2.00 89 70 86.5 188.3M04 3.00 2.45 120 71 87.0 209.0M05 2.88 2.39 87 48 87.6 212.5



4. Wood filled plastics: the time has come! Plastics Engineering2005.

5. Kosonen, M. L.; Wang, B.; Caneba, G. T.; Gardner, D. J.; Rials,G. Clean Prod Processes 2000, 2, 117.

6. Li, T. Q.; Wolcott, M. P. Compos Part A: Appl Sci Manufactur-ing 2004, 35, 303.

7. Devi, R. R.; Maji, T. K.; Banerjee, A. N. J Appl Polym Sci 2004,93, 1938.

8. Baysal, E.; Ozaki, S. K.; Yalinkilic, M. Wood Sci Technol 2004,38, 405.

9. Georgopoulos, S. T.; Tarantili, P.; Avgerinos, E.; Andreopou-los, A.; Koukios, E. Polym Degrad Stab 2005, 90, 303.

10. Jiang, H.; Kamdem, D. P. J Vinyl Additive Technol 2004, 10, 70.11. Shah, B. L.; Matuana, L. M.; Heiden, P. A. J Vinyl Additive

Technol 2005, 11, 160.12. Kim, J.; Peck, J. H.; Hwang, S.; Hong, J.; Hong, S. C.; Huh, W.;

Lee, S. J Appl Polym Sci 2008, 108, 2654.13. Zhao, Y.; Wang, K.; Zhu, F.; Xue, P.; Jia, M. Polym Degrad

Stab 2006, 91, 2874.14. Pickering, K. L.; Abdalla, A.; Ji, C.; Mcdonald, A. G.; Franich,

R. A. Compos Part A: Appl Sci Manufacturing 2003, 34, 915.15. Valadez-Gonzalez, A.; Cervantes-Uc, J. M.; Olayo, R.; Herrera-

Franco, P. J. Compos Part B: Eng 1999, 30, 321.16. Torry, S.; Campbell, A.; Cunliffe, A.; Tod, D. Int J Adhes

Adhesives 2006, 26, 40.17. Wen, J.; Wilkes, G. L. Polym Bull 1996, 37, 51.18. Bengtsson, M.; Oksman, K. Compos Sci Technol 2006, 66,

2177.19. Kuan, C.-F.; Ma, C.-C. M.; Kuan, H.-C.; Wu, H.-L.; Liao, Y.-M.

Compos Sci Technol 2006, 66, 2231.20. Percec, V.; Popov, A. V.; Ramirez-Castillo, E.; Weichold, O. J

Polym Sci Part A: Polym Chem 2003, 41, 3283.21. Percec, V.; Popov, A. V.; Ramirez-Castillo, E.; Coelho, J. F. J.;

Hinojosa-Falcon, L. A. J Polym Sci Part A: Polym Chem 2004,42, 6267.

22. Percec, V.; Popov, A. V.; Ramirez-Castillo, E.; Coelho, J. F. J. JPolym Sci Part A: Polym Chem 2005, 43, 773.

23. Percec, V.; Popov, A. V. J Polym Sci Part A: Polym Chem2005, 43, 1255.

24. Qiu, J.; Charleux, B.; Matyjaszewski, K. Prog Polym Sci 2001,26, 2083.

25. Bledzki, A. K.; Faruk, O. Compos Part A: Appl Sci Manufac-turing 2006, 37, 1358.

26. Kuan, C.-F.; Kuan, H.-C.; Ma, C.-C. M.; Huang, C.-M. ComposPart A: Appl Sci Manufacturing 2006, 37, 1696.

27. Sailaja, R. Compos Sci Technol 2006, 66, 2039.28. Coelho, J. F. J.; Silva, A. M. F. P.; Popov, A. V.; Percec, V.;

Abreu, M. V.; Goncalves, P. M. O. F.; Gil, M. H. J Polym SciPart A: Polym Chem 2006, 44, 2809.

29. Coelho, J. F. J.; Silva, A. M. F. P.; Popov, A. V.; Percec, V.;Abreu, M. V.; Goncalves, P. M. O. F.; Gil, M. H. J Polym SciPart A: Polym Chem 2006, 44, 3001.

30. Coelho, J. F.; Carreira, M.; Popov, A. V.; Goncalves, P. M.; Gil,M. Eur Polym J 2006, 42, 2313.

31. Coelho, J. F.; Carreira, M.; Goncalves, P. M.; Popov, A. V.; Gil,M. H. J Vinyl Additive Technol 2006, 12, 156.

32. Coelho, J. F. J. New Technologies for Homopolymerizationand Copolymerization of Vinyl Chloride, Ph.D. Thesis, Uni-versity of Coimbra, 2006.

33. Britcher, L. G.; Kehoe, D. C.; Matisons, J. G.; Swincer, A. G.Macromolecules 1995, 28, 3110.

34. Tavares, M. I. B.; Monteiro, E. E. C. Polym Test 1995, 14, 273.35. Prasath, R. A.; Nanjundan, S.; Pakula, T.; Klapper, M. J Appl

Polym Sci 2006, 100, 1720.36. Herve, M.; Hirschinger, J.; Granger, P.; Gilard, P.; Deflandre,

A.; Goetz, N. Biochim Biophys Acta 1994, 1204, 19.37. Wilberg, K. B.; Pratt, W. E.; Bailey, W. F. J Org Chem 1980, 45,

4936.38. Menard, K. P. Dynamic Mechanical Analysis - A Practical

Introduction, 1st ed.; CRC Press: London, 1999.

2738 ROCHA ET AL.


poly(vinyl chloride) and wood flour press mould composites: new bonding strategies

Documents